What happened when Anthropic's Claude AI ran a small shop for a month

A mini-fridge stocked with drinks and snack baskets might seem like an ordinary office setup, but for one month, it became the testing ground for an AI-powered vending machine manager named Claudius. Anthropic, the AI company behind the experiment, aimed to assess how well current AI systems could handle real-world business tasks.

The results were a mixed bag. Claudius showed promise in some areas, like using web searches to find specialty items. However, it also made puzzling mistakes, including hallucinating conversations with non-existent employees and claiming to visit the fictional home of The Simpsons.

Despite these hiccups, Anthropic sees potential for AI in business management. As their blog post notes, "AI middle-managers are plausibly on the horizon. It's worth remembering that the AI won't have to be perfect to be adopted; it will just have to be competitive with human performance at a lower cost in some cases."

This experiment offers a glimpse into the evolving landscape of AI in the workplace, highlighting both its current limitations and future possibilities.